Hello everyone,When I used processor 2 2 1,I found it spend more time than one CPU,What can be the reason for this? Does anyone know?
Attachment | Size |
---|---|
![]() | 704.01 KB |
![]() | 57.42 KB |
A project by DCS Computing and CFDEMresearch
Hello everyone,When I used processor 2 2 1,I found it spend more time than one CPU,What can be the reason for this? Does anyone know?
Attachment | Size |
---|---|
![]() | 704.01 KB |
![]() | 57.42 KB |
This website uses cookies to ensure you get the best experience on our website. By continuing to browse the site you are agreeing to our use of cookies. This website uses cookies. By continuing to browse the site you are agreeing to our use of cookies. More info
Philippe | Mon, 04/10/2017 - 12:25
system too small
My initial guess would be that your system is too small. In a system with 200 particles, the parallelization overhead is very likely to completely destroy your performance. Running a case in parallel only makes sense if you have at least several thousand particles per processor.
Bella | Mon, 04/10/2017 - 13:33
Thanks,but not work
Thanks for your reply,but I try 13705 atoms,it's still the same.loop time of 322.646 on 4 procs for 80000 steps with 13705 atoms
Loop time of 198.354 on 1 procs for 80000 steps with 13705 atoms.
mschramm | Mon, 04/10/2017 - 16:46
Communication Overhead?
Were you running one of the example cases? Which one?
What happens if you use a different processor arrangement? 1 2 2? 2 1 2? Letting LIGGGHTS choose?
Bella | Wed, 04/12/2017 - 10:24
The same phenomenon
Not example case,different processor arrangement is the same time